Automatic Piano Melody Generation Method Integrating Note Starting Point Detection and Multiple Fundamental Frequency Estimation Algorithm

Harmony is a combination of multiple notes which play an important role in melodic composition for piano. Note onset detection as well as multiple fundamental frequency estimation helps to accurately determine note onset times and add appropriate harmonic effects to melodic monophony. The study introduces a note onset detection method based on tonal filters. To add harmonic effects to the melodic monotones and enable the automatic creation of excellent piano melodies, the principal note and harmonic components derived from multi-base frequency estimation are also utilized. The results show that the note starting point detection method proposed in the study can significantly highlight the peaks with small differences in the amplitude of individual peaks. In the meantime, the multiple fundamental frequency estimate method's precision, recall, and F1 value reach high levels of 89.69%, 91.36%, and 90.27%, respectively, when the total combined notes is 2. Taken together, the proposed method of the study has significant performance advantages and provides reliable technical support for piano melody composition.